flow-next-audit

You can identify outdated knowledge entries by running an audit against the codebase to classify entries as Keep, Update, Consolidate, Replace, or Delete. The process validates frontmatter fields like track, category, module, tags, and status.

Yes, you can automatically fix conflicting memory entries without manual prompts by using the autofix mode. This mode applies safe changes automatically and generates a complete report of all modifications made to the memory entries.

A memory audit checks codebase knowledge files by validating frontmatter fields like track, category, module, tags, and status. It investigates entries across bug and knowledge tracks to collect evidence and determine if guidance still reflects the current codebase.

You consolidate conflicting guidance entries by running an audit that uses phase-driven investigation to detect contradictions. The audit provides actionable recommendations to Consolidate or Replace entries, applying changes via the flowctl workflow with proper citations.

A limitation of automated memory entry audits is that ambiguous entries cannot be automatically resolved; they must be marked as stale. Additionally, the audit relies on the flowctl workflow to apply changes, requiring proper citation setup to function correctly.